At that time we had started to collect these oddball IO methods:
loadintovector
- PetscErrorCode (*loadintovectornative)(PetscViewer,Vec);
- PetscErrorCode (*viewnative)(Vec,PetscViewer);
and Vec/MatLoad() actually created the object and load into it.
I wanted XXXView/XXXLoad() to take an already created XXXX (allowing the
user to determine type etc) and get rid of all the oddball methods. I think
the lose of all the native support was accidental, hence Shri put some of it
back in. If there is more we should have that is not there then we should add
it back in.
Barry
On Aug 8, 2013, at 9:51 AM, Shri <[email protected]> wrote:
> I was fairly new to PETSc during this time and I guess Barry (or Hong?) asked
> me to remove the support for PETSC_VIEWER_NATIVE. I don't recall any
> rationale that I had for removing this support and then putting it back
> later. From the commits, I see that I was working on updating MatLoad() and
> VecLoad() during this time.
>
> Shri
>
> On Aug 8, 2013, at 9:17 AM, Jed Brown wrote:
>
>> Shri, quite a while back now, you removed a bunch of support for
>> PETSC_VIEWER_NATIVE:
>>
>> https://bitbucket.org/petsc/petsc/commits/742cafdd48f845c927348e65bdee73999c739268
>>
>> and then added some of it back
>>
>> https://bitbucket.org/petsc/petsc/commits/f440316591b37bf1119cb52801fe97b42a9e8902
>>
>> Since this commit, there is no way to view vectors in native format and
>> only MPIDENSE matrices that can be viewed in native format. All other
>> cases ignore the format, and in particular, DMDA vectors cannot be
>> viewed in PETSc ordering.
>>
>> This is a big behavioral change that was not documented and makes
>> debugging certain things a hell of a lot harder. We have had a number
>> of flawed petsc-dev and petsc-maint threads in which we gave advice to
>> use PETSC_VIEWER_NATIVE, but that advice would have been wrong since
>> PETSC_VIEWER_NATIVE has done nothing for years now.
>>
>> What was the rationale behind this change?
>>
>>
>> Note this thread from a week before the commits above, though
>> PETSC_VIEWER_NATIVE was being used for a totally different purpose.
>>
>> http://mail-archive.com/search?l=mid&[email protected]
>
